duke law school legal studies research paper …...other central banks similarly have been cutting...

DUKE LAW SCHOOL

Duke Law School Legal Studies Research Paper Series

Research Paper No. 175 November 2007

Protecting Financial Markets: Lessons from the Subprime Mortgage Meltdown

Steven L. Schwarcz Stanley A. Star Professor of Law and Business

Founding/Co-Academic Director, Duke Global Capital Markets Center Duke University

[email protected]

Copyright 2007 by Steven L. Schwarcz

mailto:[email protected]

2/19/08 ddrraaff

tt

Protecting Financial Markets: Lessons from the Subprime Mortgage Meltdown 1

Steven L. Schwarcz2

Abstract: Why did the recent subprime mortgage meltdown undermine financial market stability notwithstanding the protections provided by market norms and financial regulation? This article attempts to answer that question by identifying anomalies and obvious protections that failed to work, and then by examining hypotheses that might explain the anomalies and failures. The resulting explanations provide critical insights into protecting financial markets.

I. INTRODUCTION.......................................................................................................... 2 II. IDENTIFYING ANOMALIES AND FAILURES....................................................... 3 III. SEARCHING FOR LESSONS.................................................................................... 7 A. If disclosure provides investors with all the information needed to assess investments, why did so many investors make poor decisions? .............................................................. 7 B. Is there something structurally wrong about how structured finance worked in the mortgage context?............................................................................................................. 15 C. Why did a problem with the subprime mortgage-backed securities markets quickly infect the markets for prime mortgage-backed securities and other asset-backed securities?.......................................................................................................................... 21 D. Why was the market-discipline approach, along with other existing protections, insufficient?....................................................................................................................... 25 E. Why did the rating agencies fail to anticipate the downgrades? ................................. 27 IV. CONCLUSIONS ....................................................................................................... 31

1 Copyright © 2008 by Steven L. Schwarcz. 2 Stanley A. Star Professor of Law & Business, Duke University School of Law; Founding/Co-Academic Director, Duke Global Capital Markets Center. E-mail: [email protected]. The author thanks Richard Bookstaber, Jonathan C. Lipson, Joseph H. Sommer, . . . and participants in faculty workshops at Temple University, James E. Beasley School of Law, Duke Law School, The University of Tennessee (College of Law, College of Business Administration, and Corporate Governance Center), . . . , and a workshop on “Structured Finance and Loan Modification” at the United States Federal Reserve Bank of Cleveland for comments. He also thanks Mark Covey for excellent research assistance.

Subprime Meltdown.doc

mailto:[email protected]

2

I. INTRODUCTION

Congress has been holding hearings on threats to the financial system in response

to the recent subprime (or sub-prime) mortgage meltdown and its impact on the

mortgage-backed, and other asset-backed, securities markets.3 Central banks worldwide

have likewise expressed concern about this crisis and its potential systemic effects. The

United States Federal Reserve Bank, for example, is attempting to reduce the likelihood

that this crisis might affect other financial markets and the economy by cutting the

discount rate, which is the interest rate the Federal Reserve charges a bank to borrow

funds when a bank is temporarily short of funds,4 and also by cutting the federal funds

rate that banks charge other banks on interbank loans.5 The European Central Bank and

other central banks similarly have been cutting the interest rate they charge to borrowing

banks.6

These steps, however, have directly impacted banks, not financial markets.7

Furthermore, changes in monetary policy, such as cutting interest rates, may not work

quickly enough—or may be too weak—to quell panics, falling prices, and the potential

for systemic collapse.8 This somewhat anachronistic focus on banks, not markets, ignores

3 See, e.g., Systemic Risk: Examining Regulators’ Ability to Respond to Threats to the Financial System: Hearing Before the H. Committee on Financial Services, 110th Cong. (Oct. 2, 2007), available at http://www.house.gov/apps/list/hearing/financialsvcs_dem/ht1002072.shtml (hereinafter Systemic Risk Hearing). 4 See Greg Ip, Robin Sidel & Randall Smith, Stronger Steps: Fed Offers Banks Loans to Ease Credit Crisis, WALL ST. J., Aug. 18-19, 2007, Weekend Ed., at A1; http://www.duke.edu/%7Echarvey/Classes/wpg/bfglosd.htm (visited Aug. 20, 2007). 5 Greg Ip, Fed’s Rate Cut Could Be Last For a While, WALL ST. J., Nov. 1, 2007, at A1; http://www.newyorkfed.org/aboutthefed/fedpoint/fed15.html (visited Nov. 25, 2007). 6 See, e.g., Randal Smith, Carrick Mollenkamp, Joellen Perry, & Greg Ip, Loosening Up: How a Panicky Day Led the Fed to Act, WALL ST. J., Aug. 20, 2007, at A8. 7 Ip, Sidel, & Smith, supra note 4, at A8 (observing that “the [Fed’s] discount window’s reach in the current crisis is limited by the fact that only banks can use it, and they aren’t the ones facing the greatest strains”). 8 Mortimer B. Zuckerman, Preventing a Panic, U.S. NEWS & WORLD REP., Feb. 11, 2008, at 64, 63 (observing that “[l]ower interest rates prompted by the Federal Reserve Bank cannot fully counter the forces of credit and liquidity contraction” caused by the


3

the ongoing trend towards disintermediation—or enabling companies to access the

ultimate source of funds, the capital markets, without going through banks or other

financial intermediaries.9 We thus are using tools to protect the financial system that have

not kept up with underlying changes in the system. In a financially disintermediated

world, the old protections may no longer be reliable.

This article explores why the subprime financial crisis occurred notwithstanding

the array of existing protections included in financial regulation, market norms and

customs, and the market-discipline approach undertaken by the second Bush

administration,10 and what this crisis can teach us about protecting financial markets.11

The article begins by identifying anomalies and obvious protections that failed to work.

The article then searches for lessons by examining various hypotheses of why these

anomalies and failures may have occurred.

II. IDENTIFYING ANOMALIES AND FAILURES

subprime mortgage crisis). Cf. Seth Carpenter & Selva Demiralp, The Liquidity Effect in the Federal Funds Market: Evidence from Daily Open Market Operations, 38 J. MONEY CREDIT & BANKING 901, 918-919 (2006) (concluding that although a change in monetary policy can begin to affect the cost of capital within a day, its full effects can take much longer); Serena Ng, Greg Ip, & Shefali Anand, Fed Fails So Far In Bid to Reassure Anxious Investors, WALL ST. J., Aug. 21, 2007, at A1. 9 Steven L. Schwarcz, Enron and the Use and Abuse of Special Purpose Entities in Corporate Structures, 70 U. CIN. L. REV. 1309, 1315 (2002). Capital markets are now the nation’s and the world’s most important sources of investment financing. See, e.g., McKinsey Global Institute, Mapping the Global Capital Markets Third Annual Report (Jan. 2007), reporting that as of the end of 2005, the value of total global financial assets, including equities, government and corporate debt securities, and bank deposits, was $140 trillion, available at http://www.mckinsey.com/mgi/publications/third_annual_report/index.asp. 10 See, e.g., Anthony W. Ryan, Assistant Secretary for Financial Markets, U.S. Department of the Treasury, Remarks before the Managed Funds Association Conference (June 11, 2007) (transcript on file with author), at 2 (discussing the market-discipline approach). 11 The term “subprime” includes both loans to borrowers of dubious creditworthiness and very large loans to otherwise creditworthy borrowers. MEGAN DORSEY & DAVID ROCKWELL, FINANCING: RESIDENTIAL REAL ESTATE 60 (8th ed. 1990).


http://www.mckinsey.com/mgi/publications/third_annual_report/index.asp

4

The following represent anomalies arising from, and protections that failed to

deter, the subprime mortgage meltdown: (A) If disclosure provides investors with all the

information they need to assess investments, why did so many investors make poor

decisions? (B) Securitization and other forms of structured finance (collectively,

“structured finance”), pursuant to which mortgage-backed and other forms of asset-

backed securities are issued, are supposed to diversify and reallocate risk to parties best

able to bear it. Is there something structurally wrong about how this worked in the

mortgage context? (C) Why did a problem with the subprime mortgage-backed securities

markets quickly infect the markets for prime mortgage-backed securities and other asset-

backed securities?12 (D) The second Bush administration expected that its market-

discipline approach would be sufficient, along with existing protections, to protect

against financial market instabilities. Why did this approach turn out to be insufficient?

(E) Why did the rating agencies fail to anticipate the downgrades?

In order to examine hypotheses of why these anomalies and failures may have

occurred, certain structured finance terminology must first be explained. The issuer of

mortgage-backed and other forms of asset-backed securities in structured finance

transactions is typically a special-purpose vehicle, or “SPV” (sometimes called a special-

purpose entity, or “SPE”). These securities are customarily categorized as MBS, ABS,

CDO, or ABS CDO.13 MBS means mortgage-backed securities, or securities whose

payment derives principally or entirely from mortgage loans owned by the SPV. ABS

means other asset-backed securities, or securities whose payment derives principally or

entirely from receivables or other financial assets—other than mortgage loans—owned

by the SPV. Industry participants refer to transactions in which SPVs issue MBS or ABS

as securitization.

The term “securitization” also technically includes CDO and ABS CDO

transactions. CDO, or “collateralized debt obligation,” securities are backed by—and thus

12 For an explanation of the types of securities involved in the subprime financial crisis, see infra notes 13-18 and accompanying text.


5

their payment derives principally or entirely from—a mixed pool of mortgage loans

and/or other receivables owned by an SPV. ABS CDO securities, in contrast, are backed

by a mixed pool of ABS and/or MBS securities owned by the SPV, and thus their

payment derives principally or entirely from the underlying mortgage loans and/or other

receivables ultimately backing those ABS and MBS securities.14 For this reason, ABS

CDO transactions are sometimes referred to as “re-securitization.”

Schematically, the distinctions among these categories can be portrayed as

follows:

Securitization:Collateralized Debt

Obligation: ABS CDO:(“re-securitization”)

Investors Investors Investors

SPV SPV SPV

SPV

issu

es

one

or m

ore

clas

ses

of

secu

ritie

s

SPV

issu

es

mul

tiple

cla

sses

(tr

anch

es) o

f se

curit

ies

SPV

issu

es

mul

tiple

cla

sses

(tr

anch

es) o

f se

curit

ies

One type of receivables

from multiple obligors

Different types of receivables from multiple obligors

Different types of ABS and/or MBS

Note: The ABS/MBS itself is backed by different types of receivables from multiple obligors

The classes, or “tranches,” of MBS, ABS, CDO, and ABS CDO securities issued

in these transactions are typically ranked by seniority of payment priority. The highest

priority class is called senior securities. In MBS and ABS transactions, lower priority

classes are called subordinated, or junior, securities. In CDO and ABS CDO transactions,

13 There are arcane variations on the CDO categories, such as CDOs “squared” or “cubed,” but these go beyond this article’s analysis.


6

lower priority classes are usually called mezzanine securities—with the lowest priority

class, which has a residual claim against the SPV, being called the equity.15

The senior and many of the subordinated classes of these securities are more

highly rated than the quality of the underlying receivables.16 For example, senior

securities issued in a CDO transaction are usually rated AAA even if the underlying

receivables consist of subprime mortgages, and senior securities issued in an ABS CDO

transaction are usually rated AAA even if none of the MBS and ABS securities

supporting the transaction are rated that high. This is accomplished by allocating cash

collections from the receivables first to pay the senior classes and thereafter to pay more

junior classes (the so-called “waterfall” of payment). In this way, the senior classes are

highly overcollateralized to take into account the possibility, indeed likelihood, of delays

and losses on collection.

The subprime financial crisis occurred because, with home prices unexpectedly

plummeting and adjustable-rate mortgage (ARM) interest rates skyrocketing,17 many

more borrowers defaulted than anticipated,18 causing collections on subprime mortgages

to plummet below the original estimates. Thus, equity and mezzanine classes of securities

14 “Synthetic” CDOs, which do not appear to be relevant to this article’s analysis, own derivative instruments, such as credit default swaps, rather than receivables, ABS, or MBS. 15 In MBS and ABS transactions, the term “equity” is not generally used because the company originating the securities (the “Originator”) usually holds, directly or indirectly, the residual claim against the SPV. 16 The equity class is generally not rated. 17 Although rate increases on ARM loans were not per se unexpected, the end of the liquidity glut made it harder for subprime borrowers to refinance into loans with lower, affordable, interest rates. See Kemba J. Dunham & Ruth Simon, Refinancing May be Harder to Enjoy, WALL ST. J., Nov. 24, 2007, at B1; Rick Brooks & Constance Mitchell Ford, The United States of Subprime, WALL ST. J., Oct 11, 2007, at A1. But cf. Ruth Simon, Rising Rates to Worsen Subprime Mess, WALL ST. J., Nov. 24, 2007, at A1 (reporting that many mortgages defaulted even before interest rates increased). 18 “Incentives and Failures in the Structured Finance Market: The Case of the Subprime Mortgage Market,” presentation by Anthony B. Sanders, Bob Herberger Arizona Heritage Chair Professor of Finance, Arizona State University, to the Federal Reserve


7

were impaired, if not wiped out, and in many cases even senior classes were impaired.19

Investors in these securities lost billions,20 creating a loss of confidence in the financial

markets.

III. SEARCHING FOR LESSONS

A. If disclosure provides investors with all the information needed to assess

investments, why did so many investors make poor decisions?

For this anomaly and failure, this article examines the following hypotheses:

Hypothesis: The disclosure was inadequate because the depth of the fall of the housing market exceeded reasonable worst-case scenarios. Mortgage loans, which were the asset class supporting the MBS as well as a significant portion of the CDO and ABS CDO securities, therefore turned out to be severely undercollateralized in many cases.

Any failure to envision the actual worst-case scenario may have reflected, to some

extent, a failure to take a sufficiently long view of risk. Some explain the near-collapse of

Long-Term Capital Management (LTCM) as resulting from that type of failure.21

Investors and other market participants looked to the recent past as an example of what

could happen to home prices,22 but they did not always look to worst-case possibilities,

such as the experience of the Great Depression.23

Bank of Cleveland at its workshop on “Structured Finance and Loan Modification,” Nov. 20, 2007 (on file with author). 19 Carrick Mollenkamp & Serena Ng, Wall Street Wizardry Amplified Credit Crisis, WALL. ST. J., Dec. 27, 2007, at A1 (reporting on the downgrade of one CDO’s triple-A rated tranches to junk status). 20 Reference in this article to “investors” means investors in capital market securities, not investors in the homes financed by the mortgage loans ultimately backing such securities. 21 See, e.g., Paul Krugman, Roshomon in Connecticut, SLATE MAG., Oct. 2, 1998. 22 Jack Guttentag, Shortsighted About the Subprime Disaster, WASH. POST, May 26, 2007, at F02 (explaining that because housing prices had been rising for a long period of time, it was assumed that they would continue to rise). 23 Christine Harper, Death of VaR Evoked as Risk-Taking Vim Meets Taleb's Black Swan, available at


8

These types of failures are inevitable, though, because “reasonable worst-case

scenarios” are judgment calls that are, necessarily, made ex ante. It does not appear

unreasonable, for example, to have viewed the Great Depression as unique.

Some failures to take a sufficiently long view of risk reflect behavioral bias due to

associations with recent similar events. Those failures are discussed separately.24

Hypothesis: The disclosure was adequate, but many investors failed to read it carefully enough or appreciate what they were reading.

This hypothesis has several possible sub-hypotheses contributing to the ultimate

failure. The first is over-reliance, insofar as investors may have relied heavily, and

perhaps in some cases exclusively, on third parties. For example, one commentator

http://www.bloomberg.com/apps/news?pid=20601109&sid=axo1oswvqx4s&refer=home (reporting that financial models at Merrill Lynch, Morgan Stanley, and UBS failed to foresee the decline in housing prices). See generally, Nassim Taleb, The Black Swan: The Impact of the Highly Improbable (2007). One commentator suggests that the disclosure also did not adequately address the relatively illiquid nature of the securities: “It is true that the level of default was unusually high, but the bulk of the problem is coming from liquidity issues—no one wants to hold these [securities], and if you try to find [a buyer] you have to trade them at a very low price.” E-mail from Richard Bookstaber, author, A DEMON OF OUR OWN DESIGN, infra note 32, to the author (Nov. 30, 2007). Lack of liquidity, however, appears to have been a standard disclosure item, such as the following disclosure taken from highlighted risk factors on p. S-28 of the March 12, 2007 Prospectus Supplement for Soundview Home Loan Trust 2007-WMC1, as Issuing Entity, Financial Asset Securities Corp., as Depositor, Countrywide Homes Loans Servicing LP, as Servicer: “There is no assurance that . . . a secondary market [in the securities] will develop or, if it develops, that it will continue. Consequently, you may not be able to sell your [securities] readily or at prices that will enable you to realize your desired yield. The market value of the [securities] are likely to fluctuate; these fluctuations may be significant and could result in significant losses to you.” I therefore believe that the problem was less failure of the illiquidity risk to be disclosed than investor failure to appreciate that disclosure. See infra notes 23-36 and accompanying text. Query, however, whether anyone knew—much less knew enough to disclose—the extent of the illiquidity problem. See e-mail from Bookstaber, supra (observing that “no one knew how levered funds were, and therefore how quickly they would need to dump [securities] if they faced a market shock”). 24 See infra notes 34-36 and accompanying text (discussing the availability heuristic).


http://www.bloomberg.com/apps/news?pid=20601109&sid=axo1oswvqx4s&refer=home

http://www.amazon.com/exec/obidos/search-handle-url/002-8658832-3689655?%5Fencoding=UTF8&search-type=ss&index=books&field-author=Nassim%20Nicholas%20Taleb

9

argues that investors over-relied on the underwriter or arranger selling them the

securities:

Investors have the prospectuses to rely on, but the reality is that they have not taken any responsibility for reading the detail of the documentation or digesting the risks involved. These investors are still under the impression that the arranger will look after their interests and are yet to appreciate the need to negotiate what are highly complicated bilateral agreements.25

Because this flies in the face of caveat emptor, it seems dubious unless the

underwriter/arranger’s interests were aligned with that of the investors.26 Those interests

were somewhat aligned, however, in ABS CDO transactions where underwriters

customarily purchased some portion of the “equity” tranches in order to demonstrate their

belief in the securities being sold. In this context, aligning the interests of sellers and

investors actually may have worked against investor caution.

Investors also may have over-relied on rating-agency ratings, without necessarily

engaging in (or at least fully performing) their own due diligence. This article later

examines why rating agencies failed to anticipate the downgrades.27 Even if investors

performed their own due diligence, agency-cost conflicts28 and lack of economy of

scale29 may have limited the extent to which they could have done a better job of

assessing creditworthiness than the rating agencies.

Another sub-hypothesis is that, as a result of a market “bubble,” “many investors,

swept up in the euphoria of the moment, failed to pay close attention to what they were

25 Daniel Andrews, The Clean Up: Investors Need Better Advice on Structured Finance Products, 26 INT’L FIN. L. REV. 14, 14 (Sept. 2007). 26 This form of the hypothesis, of course, is now even more dubious as a predictor of (at least near-term) future investor reliance. 27 See Part III.E, infra. 28 See infra notes 43-47 and accompanying text. 29 Individual investors face relatively high costs to assess the creditworthiness of complex ABS, ABS, CDO, and ABS CDO securities, whereas rating agencies make this assessment on behalf of many individual investors, thereby achieving an economy of scale.


10

buying.”30 Bubbles can start quite easily. If, for example, a particular stock unexpectedly

gains in value, the losers (e.g., those shorting the stock) will tend to withdraw from that

market and the winners will tend to increase their investment, driving up the price even

further. Soon other winners are attracted to the stock and other losers cut their losses and

stop shorting the stock. This process is aided by almost inevitable explanations of why it

is “rational” for the price to keep going up and why the traditional relationship of price to

earnings does not apply. Even investors who recognize the bubble as irrational may buy

in, hoping to sell at the height of the bubble before it bursts.31 In these ways, price

movements can become somewhat self-sustaining.32

Compare the tulip “bubble” in 17th century Holland, in which certain tulips were

highly prized and their bulbs were sold for thousands of guilder. Almost everyone got

caught up in the excitement of buying and selling tulip bulbs, usually on credit and with

the intention of making a quick profit, but many who speculated on credit were left with

crushing debts when the market finally crashed.33

Occasional bubbles may well be an inevitable side effect of a market economy.

A third sub-hypothesis is bounded rationality. Bubbles do not necessarily require

individual investors to behave irrationally. In contrast, investors can make poor decisions,

notwithstanding disclosure, because of their bounded rationality. There are at least two

ways in which this can occur. To some extent, investor failure in the subprime financial

crisis may have resulted from herd behavior.34 To some extent also, it may have resulted

30 Alan S. Blinder, Six Fingers of Blame in the Mortgage Mess, N.Y. TIMES, Sept. 30, 2007, at BU 4. 31 Sam Segal, Tulips Portrayed: The Tulip Trade in Holland in the 17th Century, in THE TULIP: A SYMBOL OF TWO NATIONS 181 (Michael Roding & Hans Theunissen eds., 1993). 32 RICHARD BOOKSTABER, A DEMON OF OUR OWN DESIGN: MARKETS, HEDGE FUNDS, AND THE PERILS OF FINANCIAL INNOVATION 169–70 (2007). 33 Segal, supra note 31, at 17–20. 34 Cf. Steven L. Schwarcz, Rethinking the Disclosure Paradigm in a World of Complexity, 2004 U. ILL. L. REV. 1, 14-15 (observing and explaining this behavior in a related context).


11

from the availability heuristic, under which people overestimate the frequency or

likelihood of an event when examples of, or associations with, similar events are easily

brought to mind. People typically overestimate the divorce rate, for example, if they can

quickly find examples of divorced friends.35 Similarly, once past financial crises recede

in memory and investors are making money, investors always “go for the gold.”36

Hypothesis: The disclosure was inherently inadequate because the transactions were so complex that many investors could not understand them.37

This hypothesis turns on the extraordinary complexity of CDO and ABS CDO

transactions. [Illustrate that complexity by discussing the 300+-page prospectuses of a

typical ABS CDO transaction.38]

This hypothesis, if true, would extend the thesis in Rethinking the Disclosure

Paradigm in a World of Complexity39 beyond investors in an Originator’s40 securities to

35 Paul Slovic, Baruch Fischhoff & Sarah Lichtenstein, Facts Versus Fears: Understanding Perceived Risk, in JUDGMENT UNDER UNCERTAINTY: HEURISTICS AND BIASES 463, 465 (Daniel Kahneman et al. eds., 1982). 36 Cf. Larry Light, Bondholder Beware: Value Subject to Change Without Notice, BUS. WK., at 34 (Mar. 29, 1993) (“[b]ondholders can—and will—fuss all they like. But the reality is, their options are limited: higher returns or better protection. Most investors will continue to go for the gold.”) (discussing, in the context of but several years after the “Marriott split,” that investors favor higher interest rates over “event risk” covenants once examples of events justifying the covenants have receded in memory, even though they could reoccur). 37 See, e.g., Credit & Blame: How Rating Firms’ Calls Fueled Subprime Mess, WALL ST. J., Aug. 15, 2007, at A1 (quoting a market observer): “A lot of institutional investors bought [mortgage-backed] securities substantially based on their ratings [without fully understanding what they bought], in part because the market has become so complex.” Cf. Blinder, supra note 30 (arguing that the MBS, especially the CDOs, “were probably too complex for anyone’s good”). See also Malcolm Gladwell, Open Secrets: Enron, Intelligence, and the Perils of Too Much Information, NEW YORKER, Jan. 8, 2007 (distinguishing between transactions that are merely “puzzles” and those that are truly “mysteries”). To the extent complexity is merely a puzzle, investment bankers theoretically could understand it. 38 [cite1] [Query also the extent, if any, to which CDO and ABS CDO disclosure was intentionally non-transparent in order to protect trade secrets. To what extent does that explain why underwriters of these securities, who typically put together the deals, needed to invest in a portion of the equity? cite1]


12

investors in an SPV’s securities. The proposal of that article nonetheless can help to

inform the analysis. That article proposes that investors in an Originator’s securities be

protected in a supplementary manner by restricting conflicts of interest in complex

transactions for which disclosure would be insufficient.41 The rationale is that, absent

conflicts, the Originator’s management will make decisions that more closely reflect the

interests of the Originator’s investors.

The same approach has potential application to investors in an SPV’s securities,

particularly when the SPV transaction is so complex (as some CDO and ABS CDO

transactions apparently were) that disclosure would be insufficient. In that context, there

are at least two ways in which material conflicts arise. For securities backed by subprime

mortgages, the interests of mortgage originators, absent their taking a prior or pari passu

risk of loss, are misaligned with that of investors in those securities.42 To mitigate this

type of conflict, perhaps mortgage originators should be required to take some risk of

loss. Secondly, agency-cost conflicts arise when the interests of individual investment

bankers who structure, sell, or invest in securities are misaligned with the interests of the

institutions for which they work.43 For example, certain losses of institutional investors

such as Bear Stearns appear to have resulted from losses in CDO investments by

controlled or managed hedge funds.44 If managers of those hedge funds were paid

according to hedge-fund industry custom—in which “fund managers reap large rewards

on the upside without a corresponding punitive downside”45—they would have had

39 Schwarcz, supra note 34. 40 The term “Originator” is defined supra note 15. 41 Schwarcz, supra note 34, at 30. See also id. at 32-33 (showing how to identify these transactions, defined as “disclosure-impaired transactions”). 42 See infra notes 54-64 and accompanying text. 43 Most investors were institutions. [cite, including Rule 144A exemptions and $500,000 investment limit for ABS CDO securities.] 44 Kate Kelly, Serena Ng & David Reilly, Two Big Funds At Bear Stearns Face Shutdown—As Rescue Plan Falters Amid Subprime Woes, Merrill Asserts Claims, WALL. ST. J., June 20, 2007, at A1. 45 James Surowiecki, Performance-Pay Perplexes, NEW YORKER, Nov. 12, 2007, at 34.


13

significant conflicts of interest with the institutions owning the hedge funds.46 To

mitigate this type of conflict, these individuals should be paid in a manner that better

aligns their interests with the interests of the institutions for which they work.

Restricting conflicts of interest, as a supplement to disclosure, is only a second-

best solution. It would not, for example, solve the problem that, even absent conflicts,

individual investment bankers might have insufficient incentives to try to completely

understand the highly complex transactions in which they recommend their institutions

invest. Such individuals might, for example, view the possibility of losses as remote, or

anticipate being in a new job if and when losses occurred, or simply feel safe following

the herd of other bankers.47

There do not, however, appear to be any perfect solutions. Government already

takes a somewhat paternalistic stance to mitigate disclosure’s inadequacy by mandating

minimum investor sophistication for investing in complex securities, yet sophisticated

investors and qualified institutional buyers (QIBs) are the very investors who lost the

most money in the subprime financial crisis.48 And any attempt by government to restrict

firms from engaging in complex transactions would be highly risky because of the

potential of inadvertently banning beneficial transactions.49

46 In this regard, the reader should distinguish these conflicts of interest not only from the agency-cost problem discussed above but also from the potential conflict of interest discussed infra notes 54-62 and accompanying text between mortgage originators and investors. 47 Schwarcz, supra note 34, at 2, 14-15. Outside of an institutional-industry context, there may be further misalignment of incentives because of higher employee turnover. Id. at 14 (observing that employee turnover reduces accountability). 48 See, e.g., Jenny Anderson, Wall St. Banks Confront a String of Write-Downs, N.Y. TIMES, Feb. 19, 2008, at C1 (reporting that “major banks . . . have already written off more than $120 billion of losses stemming from bad mortgage-related investments”); Randall Smith, Merrill’s $5 Billion Bath Bares Deeper Divide—After Big Write-Down Tied to Mortgage Debt, O’Neal Asserts Control, WALL ST. J., Oct. 6, 2007, at A1 (reporting a total of $20 billion in write-downs by large investment banks). 49 Cf. infra note 56 and accompanying text (cautioning against “throwing out the baby with the bathwater”). Although otherwise beyond this article’s scope (see supra note 13), certain CDO products, the so-called CDOs “squared” and “cubed,” might be worthy of special consideration because they are subject to “cliff risk,” or suddenly losing 100% of


14

Hypothesis: Even when disclosure is adequate and investors understand it perfectly (i.e., they have perfect knowledge of the risk), disclosure alone will be inadequate to address at least systemic risk in financial markets.

Systemic risk is the risk that (i) an economic shock such as market or institutional

failure triggers (through a panic or otherwise) either (x) the failure of a chain of markets

or institutions or (y) a chain of significant losses to financial institutions, (ii) resulting in

increases in the cost of capital or decreases in its availability, often evidenced by

substantial financial-market price volatility.50 Disclosure alone will be inadequate to

prevent systemic risk because, like a tragedy of the commons, the benefits of exploiting

finite capital resources accrue to individual market participants, each of whom is

motivated to maximize use of the resource, whereas the costs of exploitation, which

affect the real economy, are distributed among an even wider class of persons.51 Investors

are therefore unlikely to care about disclosure to the extent it pertains to systemic risk.

Should disclosure therefore be supplemented to address systemic risk? I address

this in a separate article,52 proposing, among other things, a liquidity provider of last

resort to purchase securities in collapsing markets in order to mitigate market instability

that would lead to systemic collapse. The liquidity provider of last resort would make its

purchases at a deep enough discount to (i) make a profit, or at least be repaid, and (ii)

mitigate moral hazard by impairing existing investors.

their value. See, e.g., Michiko Whetten & Mark Adelson, Nomura Fixed Income Research, CDOs-Squared Demystified 12-13 (Feb. 4, 2005); Leverage and Junk Science: A Credit Crunch Cocktail, TOTAL SECURITIZATION, Sep. 20, 2007. In this context, the tort law doctrine of “unavoidably unsafe products” may help to inform a regulatory analysis. In tort law, an “unavoidably unsafe product” is subject to strict liability unless its utility outweighs its risk. Joanne Rhoton Galbreath, Products liability: What Is an “Unavoidably Unsafe” Product, 70 A.L.R.4th 16, §3 (1989). For example, the vaccine for rabies is inherently dangerous, but rabies can result in death so the vaccine is not subject to strict liability. RESTATEMENT (SECOND) OF TORTS § 402A cmt. k (1965). 50 See Schwarcz, infra note 52. 51 In other words, the externalities of systemic failure include social costs that can extend far beyond market participants. 52 Steven L. Schwarcz, Systemic Risk (working paper, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1008326).


15

Summary. The discussion above suggests that multiple causes, viewed

collectively, help to explain why so many investors make poor investment decisions

notwithstanding disclosure. Although there do not appear to be empirical ways to test the

validity of these explanations, some investors may have taken too brief a view of risk in

the housing market or have been swayed by the fact that, in recent memory, home prices

had only been rising. Some investors may have simply followed the herd in their

investments, while others—possibly recognizing the bubble forming in the market for

CDO and ABS CDO securities—may have invested anyway, hoping prices would

continue to rise and their investments would rise in value. Investors also may have relied

excessively on credit ratings, without performing their own due diligence. In the case of

investments in ABS CDO transactions, investors additionally may have over-relied on the

judgment of underwriters who had purchased portions of the “equity” tranches. Finally,

certain of the CDO and ABS CDO transactions may have been so complex that

disclosure was inherently inadequate.

B. Is there something structurally wrong about how structured finance worked in

the mortgage context?

For this anomaly, this article examines the following hypotheses:

Hypothesis: Structured finance facilitated an easy-entrant and undisciplined mortgage lending industry by enabling mortgage lenders to sell off loans as they are made (“originate and distribute”). This created moral hazard to the extent mortgage lenders therefore did not have to live with the credit consequences of their loans. For that reason, which was probably exacerbated by the fact that they could make money on the volume of loans originated,53 the underwriting standards of mortgage lenders naturally fell.54

53 This may have been further exacerbated by the fact that certain mortgage lenders, without balance-sheet assets, simply advanced to borrowers the proceeds of selling the loans (“dry funding”). [cite] 54 See, e.g., http://www.federalreserve.gov/newsevents/testimony/bernanke20070920a.htm. There is also speculation that some mortgage-loan originators might have engaged in fraud by manipulating borrower income, and that some borrowers may have engaged in fraud by


16

Anecdotal evidence suggests this hypothesis has at least some truth.55 One

solution would be to limit the originate-and-distribute model, but that would be like

“throwing out the baby with the bathwater.” An originate-and-distribute model is critical

to the underlying funding liquidity of banks56 as well as many corporations.57

A better solution, already discussed, would be to require mortgage lenders and

other Originators to retain a risk of loss.58 In many non-mortgage securitization

transactions, for example, it is customary for Originators to bear a direct risk of loss by

overcollateralizing the receivables sold to the SPV.59 This is not always done in mortgage

securitization because mortgage loans are inherently overcollateralized by the value of

the real-estate collateral (and thus investors can effectively be overcollateralized even if

the Originator bears no risk of loss).60 It needs to be done, however, to mitigate moral

hazard.

lying about their income, in each case to qualify borrowers for loans. See, e.g., Vikas Bajaj, A Cross-Country Blame Game, N. Y. TIMES, May 8, 2007, at C1. If such fraud occurred, it would exacerbate but is unlikely to be significant enough to have caused the subprime financial crisis. 55 To some extent the drop in underwriting standards under the originate-and-distribute model may reflect distortions caused by the recent liquidity glut, in which lenders competed aggressively for business and allowed otherwise defaulting borrowers to refinance. See Ravi Balakrishnan et al., Globalization, Gluts, Innovation or Irrationality: What Explains the Easy Financing of the U.S. Current Account Deficit? 12 (Int’l Monetary Fund, IMF Working Paper No. 07/160, July 2007) (discussing this liquidity glut). 56 See, e.g., Joseph R. Mason, “Mortgage Loan Modification: Promises and Pitfalls” (undated Powerpoint presentation to the Federal Reserve Bank of Cleveland at its workshop on “Structured Finance and Loan Modification,” Nov. 20, 2007) (showing that 58% of mortgage liquidity in the United States, and 75% of mortgage liquidity in California, has come from structured finance). 57 [cite] [Consider also the theory that the liquidity provided by an originate-and-distribute model would more than compensate for a departure from normal underwriting standards. cite1] 58 See supra notes 42-43 and accompanying text. 59 Vincent Ryan, Debt in Disguise, CFO MAGAZINE, Nov. 2007 (reporting that most securitization agreements include overcollateralization). 60 [cite]


17

Some investors have taken comfort by the limited risk of loss imposed on

mortgage originators through representations and warranties.61 Representations and

warranties, however, are not always effective, becoming illusory for mortgage originators

that are unable, as in the current subprime mortgage meltdown, to pay damages for

breach.62 Prudent investors should insist that mortgage originators retain some direct risk

of loss to mitigate moral hazard.63 For this same reason, for example, banks buying loan

participations insist that the bank originating the loan retain a minimum portion, typically

at least [ten] per cent of the loan exposure, even if the loan itself is overcollateralized.64

Another possible solution is to regulate the loan underwriting standards applicable

to mortgage lenders. This approach would be akin to the Federal margin regulations G,

U, T, and X imposed in response to the 1929 stock market crash.65 The then-falling stock

values caused margin loans—that is, loans to purchase publicly-listed, or margin, stock—

to become undercollateralized, in turn causing bank lenders to fail. To protect against a

repeat of this problem, the margin regulations require margin lenders to maintain two-to-

one overcollateralization when securing their loans by margin stock that has been

purchased, directly or indirectly, with the loan proceeds.66 Imposing a minimum real-

estate-value-to-loan overcollateralization on all mortgage loans secured by the real estate

financed would likewise protect against a repeat of the subprime mortgage problem.

Unfortunately, though, it would have a high price, potentially impeding and increasing

61 [cite to Sanders/Mason] 62 Cf. Sanders, supra note 18 (arguing that mortgage originators be required to post capital, to backstop their representations and warranties, for loans originated and then sold). Representations and warranties are even more patently illusory for mortgage originators engaging in dry funding. See supra note 53 (discussing dry funding, in which mortgage originators lacking assets simply advance to borrowers the proceeds of selling the loans). 63 [Consider what impact, if any, this would have on the loan origination market. cite1] 64 [cite] Cf. Blinder, supra note 30 (suggesting that mortgage loan originators “retain a share of each mortgage”). Also cf. supra note 25 and accompanying text (discussing underwriters retaining a portion of the equity when selling ABS CDO securities). 65 Cf. Blinder, supra note 30 (suggesting a “suitability standard” for selling mortgage products and that all mortgage lenders be placed under federal regulation). 66 12 C.F.R. § 221.


18

the cost of home ownership and imposing an administrative burden on lenders and

government monitors.67

Hypothesis: Structured finance dispersed subprime mortgage risk so widely that there was no clear incentive for any given investor to monitor it.

Structured finance generally diversifies and reallocates risk, which is normally

salutary.68 Might it have excessively dispersed subprime risk?69

If this hypothesis is true, it would call into question whether incentives should be

better aligned to promote monitoring, for example by limiting the degree of risk

dispersion. To some extent this article already proposes a variant on that approach, by

67 [Also consider imposing lending standards and predatory-lending restrictions in other ways, such as North Carolina’s Home Loan Protection Act which, among other things, mandates that lenders verify borrower income and also review the borrower’s ability to repay the loan after introductory rates adjust upwards. N.C. Gen. Stat. § 24-1.1E (amended by 2007 N.C. Sess. Laws 352), and compare proposed H.R. 3915, Mortgage Reform and Anti-Predatory Lending Act of 2007, under consideration by the House Financial Services Committee. Apparently, the N.C. law has not negatively impacted home ownership. Nanette Byrnes, These Tough Lending Laws Could Travel, BUS. WK., Nov. 5, 2007, at 70 (reporting that North Carolina’s housing market has not, according to “academic studies,” been negatively impacted). See http://www.planning.unc.edu/pdf/CC_NC_Anti_Predatory_Law_Impact.pdf. But cf. Byrnes, supra (reporting that groups such as the Mortgage Bankers Association argue that tough loan underwriting standards will prevent needy borrowers from obtaining mortgage loans). But some argue that the “borrowers are not victims of inappropriate loan prospecting (such as predatory lending). Rather, they [or, at least, many] were willful participants.” Sanders, supra note 18. But cf. Gretchen Morgenson, Blame the Borrowers? Not So Fast, N.Y. TIMES, Nov. 25, 2007, at BU 1.] 68 Douglas Elmendorf, Notes on Policy Responses to the Subprime Mortgage Unraveling, The Brookings Institutions, footnote 6, available at http://www.brookings.edu/~/media/Files/rc/papers/2007/09subprimemortgageunravelling/09useconomics_elmendorf.pdf (visited Dec. 27, 2007). 69 The very assumption that structured finance reallocates risk to parties best able to bear it also may have failed in the subprime context. See, e.g., e-mail from Bookstaber, supra note 23 (indicating that “[r]ather than spreading the risk to those who were most comfortable holding the assets and taking the risk, many of the [holders] were ‘hot money’ hedge funds that would have to run for cover at the very time the risk taking function was most critical”).


http://www.brookings.edu/%7E/media/Files/rc/papers/2007/09subprimemortgageunravelling/09useconomics_elmendorf.pdf

http://www.brookings.edu/%7E/media/Files/rc/papers/2007/09subprimemortgageunravelling/09useconomics_elmendorf.pdf

19

suggesting that loan originators in an originate-and-distribute model retain some

minimum percentage or amount of risk.70

Hypothesis: Structured finance can make it difficult to work out problems with an underlying asset class—in this case, for example, making it difficult to work out the underlying mortgage loans because the beneficial owners of the loans are no longer the mortgage lenders but a broad universe of financial-market investors. As a result, mortgage defaults result in unnecessarily high losses.

News stories observe that homeowners have been unable to restructure, or

“modify,” their loans because they cannot identify who owns the loans.71 Laws

protecting mortgage borrowers, however, suggest this concern may be overstated. For

example, “[u]pon written request by the obligor, the servicer shall provide the obligor, to

the best knowledge of the servicer, with the name, address, and telephone number of the

owner of the obligation or the master servicer of the obligation.”72

In theory, servicers bridge the gap between beneficial owners of the loans and the

mortgage lenders. It is typical, for example, for originators of securitized mortgage loans,

or a specialized servicing company such as Countrywide Home Loans Servicing LP, to

act as the servicer for a fee.73 In this capacity, the servicer ordinarily retains power to

70 See supra note 64 and accompanying text (proposing this to mitigate moral hazard). 71 Gretchen Morgenson, More Home Foreclosures Loom as Owners Face Mortgage Maze, N.Y. TIMES, Aug. 6, 2007, at A1. A somewhat related issue is that, at least heretofore, individual borrowers cannot use Chapter 13 bankruptcy to restructure their home mortgage loan liabilities. See 11 U.S.C. § 1322(b)(2) & (b)(5). [Examine efforts by Congress to amend this. cite] In a corporate reorganization context, however, debtors can, with the lender’s consent, use bankruptcy to restructure their secured-loan liabilities. Compare 11 U.S.C. § 1123(a)(5) with 11 U.S.C. §§ 1126 (c) & 1129(a)(7) & (a)(8). 72 15 U.S.C. § 1641(f)(2). Identification would be even less of a problem if the underlying receivables are not consumer assets, like mortgage loans, since the amounts involved in consumer receivables are typically relatively small. 73 JAMES A. ROSENTHAL & JUAN M. OCAMPO, SECURITIZATION OF CREDIT: INSIDE THE NEW TECHNOLOGY OF FINANCE 49-51 (1988) (explaining the general structure of a grantor trust when the originator of asset-backed securities services the pool of assets); Gretchen Morgenson, Countrywide Is Upbeat Despite Loss, N.Y. TIMES, Oct. 27, 2007, at C1 (reporting that Countrywide is the nation’s largest loan servicer). A specialized collateral manager is also often appointed in ABS CDO transactions, but this collateral manager does not appear to service the underlying assets. [cite]


20

restructure the underlying loans, so long as restructuring changes are “in the best

interests” of the investors holding the securities.74 Subject to that constraint, the servicer

may even change the rate of interest, the principal amount of the loan, or the maturity

dates of the loan if, for example, the loan is in default or, in the servicer’s judgment,

default is reasonably foreseeable,75 or if the borrower is delinquent for twenty days.76

In practice, though, even when a servicer has the power to restructure a mortgage

loan and restructuring is in the best interests of investors, the servicer may be reluctant to

engage in restructuring if there is uncertainty that the transaction will generate sufficient

excess cash flow to reimburse the servicer’s costs.77 A mortgage loan servicer, for

example, “[m]ust “spend $750-$1000 to do a [loan] mod[ification] [and] can’t charge the

borrower . . . or [ if there is insufficient excess cash] the securitization trust. . . . [B]y

contrast, all foreclosure costs are reimbursed.”78 Servicers also may sometimes prefer

foreclosure over restructuring because the former is more ministerial and thus has lower

litigation risk.79

74 Morgenson, supra note 71. Sometimes, however, the servicer is limited as to the percentage of loans in a given pool that can be restructured. Morgenson, supra note 71 (observing that a servicer might, for example, be permitted to restructure only 5% of the loans). 75 This example is taken from § 3.01, at 88, of the Pooling and Servicing Agreement dated as of March 1, 2007, among Financial Asset Securities Corp., as Depositor, Countrywide Homes Loans Servicing LP, as Servicer, and Deutsche Bank National Trust Company, as Trustee, relating to Soundview Home Loan Trust Asset-Backed Certificates, Series 2007-WMC1. 76 [cite] 77 Presentation by Joseph R. Mason, Associate Professor of Finance & LeBow Research Fellow, LeBow College of Business, Drexel University, to the Federal Reserve Bank of Cleveland at its workshop on “Structured Finance and Loan Modification,” Nov. 20, 2007 (notes on Mason’s presentation on file with author) (observing that servicers will prefer to foreclose, even if it is not the best remedy, when foreclosure costs but not modification costs are reimbursed). 78 Mason, supra note 56. 79 Presentation by Kathleen C. Engel, Associate Professor of Law, Cleveland-Marshall College of Law, to the Federal Reserve Bank of Cleveland, Nov. 20, 2007 (notes on this presentation on file with author).


21

Summary. The discussion above indicates there is little structurally wrong about

how structured finance worked in the mortgage context. Although the originate-and-

distribute model of structured finance may have created a degree of moral hazard, the

model is critical to underlying funding liquidity. Moreover, the moral hazard cost can be

mitigated if, as likely will occur in the future, investors learn from the subprime crisis and

require mortgage originators to retain a direct risk of loss that goes beyond the sometimes

illusory risk borne through representations and warranties.

Structured finance can make it more difficult to work out problems with the

underlying financial assets, in this case mortgage loans, but the increased difficulty may

be able to be managed. Parties should consider, for example, writing underlying deal

documentation that sets clearer and more flexible guidelines, and ideally more certain

reimbursement procedures, for loan restructuring—especially when such restructuring is

superior to foreclosure.80 Investors (and servicers) should prefer foreclosure to

restructuring if restructuring merely delays an inevitable foreclosure.81

There nonetheless might be a residual structural concern insofar as structured

finance may have dispersed subprime mortgage risk so widely that there is no clear

incentive for any given investor to monitor the risk. Whether that has occurred is

uncertain. Even if it has, the evil is not so much risk dispersion per se as the failure to

align incentives sufficiently to promote monitoring.

C. Why did a problem with the subprime mortgage-backed securities markets

quickly infect the markets for prime mortgage-backed securities and other asset-backed

securities?82

80 In the current subprime crisis, of course, the underlying deal documentation is already in place. Because existing documentation cannot be easily renegotiated, the government might consider legislating changes. Any such changes that are subsidized in whole or part by government, however, could foster moral hazard, potentially making future homeowners more willing to take risks when borrowing. 81 Engel, supra note 79. 82 Cf. Andrews, supra note 25, at 15 (observing from the subprime financial crisis that “liquidity in markets for structured investments can disappear immediately as soon as


22

Understanding this anomaly can help to expand an understanding of how market

risk can become systemic. For this anomaly, this article examines the following

hypotheses:

Hypothesis: The MBS, ABS, CDO, and ABS CDO markets are inherently tightly coupled, both within and among such markets.

By “tight coupling,” I mean, with thanks to Rick Bookstaber, the tendency for

financial markets to move rapidly into a crisis mode with little time or opportunity to

intervene.83 Tight coupling could result from various mechanisms, even as elementary as

investor panic, guilt-by-association, or loss of confidence.84 In the subprime crisis, once

investors realized that highly-rated sub-prime mortgage-backed securities could lose

money, they began shunning all complex securitization products.85 This explanation

appears to have had particular force for asset-backed commercial paper—not

surprisingly, since commercial paper is effectively a substitute for cash (albeit one that

yields a return). Investor reaction also may have been magnified by the dramatic shift

away from the liquidity glut of the past few years, which had obscured the problem of

defaults by enabling defaulting borrowers to refinance with ease.86

there are any shocks—no buying or selling at all in an entire sector”—although not explaining why this occurred). A somewhat related question might be why the U.S. domestic real estate collapse is having a significant impact overseas. The answer is that a significant amount of the CDO and ABS CDO securities backed (directly or indirectly) by such real estate was purchased by foreign investors. [cite] 83 Systemic Risk: Examining Regulators’ Ability to Respond to Threats to the Financial System: Hearing Before the H. Committee on Financial Services, 110th Cong. (Oct. 2, 2007), available at http://www.house.gov/apps/list/hearing/financialsvcs_dem/ht1002072.shtml (statement of Richard Bookstaber). Bookstaber himself borrows this term from engineering nomenclature. 84 See, e.g., A Flight to Simplicity, FT.com, Oct. 21, 2007. 85 Cf. Markus Brunnermeier, “2007 Liquidity Crisis” (on file with author) (speculating that when investors realized how difficult it was to value mortgage structured products, the volatility of all structured products increased). 86 Cf. supra note 55 and accompanying text.


23

Tight coupling also may have been caused by adverse selection: investors were no

longer sure which securitization investments or counterparties were good and which were

bad (CDO and ABS CDO products being especially difficult to value87), so they stopped

investing in all securitization products.88 Incongruously, adverse selection may have been

made worse by the otherwise salutary effect of securitization to disperse risk: investors

were unable, in part exacerbated by the indirect holding system for securities under which

third parties cannot readily determine who ultimately owns specific securities,89 to

ascertain to whom the risk was dispersed.

Hypothesis: The MBS, ABS, CDO, and ABS CDO markets are not inherently tightly coupled, but tight coupling resulted from convergence in hedge-fund quantitatively-constructed investment strategies.90

Professors Khandani and Lo hypothesize, for example, that when a number of

hedge funds experienced unprecedented losses during the week of August 6, 2007, they

rapidly unwound sizable portfolios, likely based on a multi-strategy fund or proprietary-

trading desk.91 These initial losses then caused further losses by triggering stop/loss and

de-leveraging policies.92 To this extent, hedge fund strategies, and not securitization or

structured finance per se, are responsible for the subprime financial crisis.

87 [cite] 88 See, e.g., Zuckerman, supra note 8 (arguing that “the credit system has been virtually frozen” because “few people even know where the liabilities and losses are concentrated”). 89 Under the indirect holding system for securities, intermediary entities hold securities on behalf of investors. Issuers of the securities generally record ownership as belonging to one or depository intermediaries, which in turn record the identities of other intermediaries, such as brokerage firms or banks, that buy interests in the securities. Those other intermediaries, in turn, record the identities of investors that buy interests in the intermediaries’ interests. See Steven L. Schwarcz, Intermediary Risk in a Global Economy, 50 DUKE L.J. 1541, 1547-48 (2001). Because of this ownership chain, there is no single location from which third parties can readily determine who ultimately owns specific securities. Id. at 1583. 90 Cf. Schwarcz, supra note 52, at 13 n. 50 & 14 (discussing the danger of converging hedge-fund investment strategies). 91 Amir Khandani & Andrew W. Lo, “What Happened to the Quants in August 2007” (Sept. 20, 2007) (SSRN working paper no. 1015987). 92 Id.


24

To what extent does this hypothesis turn, however, on CDO and ABS CDO

securities being mark-to-model, not mark-to-market (because such securities are not

actively traded, there is no established market price to which to mark them)? If shared

models are wrong, an unanticipated error is shared by everyone.

Summary. The discussion above provides three explanations for why a problem

with the subprime mortgage-backed securities markets quickly infected the prime

markets.93 Faced for the first time with the reality that highly-rated tranches of sub-prime

MBS could lose money, investors appear to have lost confidence, shunning all complex

securitization products. To this extent, future investors should try to better understand

these types of investments so that confidence is built on a firmer foundation.

Adverse selection also helps to explain the rapid infection. Investors became

uncertain which securitization products, and indeed which securitization counterparties,

were good and which were bad. They therefore stopped investing in all securitization

products. Adverse selection can be mitigated through information, in this case, for

example, by valuing the securities and ascertaining the holdings of securitization

counterparties. Because there was no market for CDO and ABS CDO securities,

however, these securities could not be valued at “market.” Valuation therefore was priced

off quantitative models. Marking-to-model, however, creates intrinsic valuation

uncertainties, and indeed the valuations priced off those models proved hopelessly

unreliable. The indirect holding system for securities also made it very difficult to

93 There also might have been amplifying mechanisms that exacerbated or expanded market losses. For example, highly leveraged hedge funds apparently borrowed money from banks and invested in significant amounts of MBS, CDO, and ABS CDO securities backed by subprime mortgages. [cite] Failure of these hedge funds resulting from losses on these securities can affect the bank lenders. Another possible amplifying mechanism is that certain bank-sponsored investment conduits purchased AAA-rated CDO and ABS CDO securities with the proceeds of short-term commercial paper. As the CDO and ABS CDO securities were marked down in value and investors failed to roll over their commercial paper, the bank sponsors faced the prospect of having to make payments to the conduits pursuant to liquidity and credit-enhancement facilities. [cite] See also infra note 117 and accompanying text.


25

ascertain whether CDO and ABS CDO securities were held by securitization

counterparties, and as long as that system continues to dominate securities holdings this

difficulty will remain.

The third explanation is also related to valuation. Absent a real market, valuation

of CDO and ABS CDO securities must, as indicated, be priced off quantitative models. It

is critical, then, that the range of models used by investors be sufficiently diverse that

errors in one model will not cut across all models.

D. Why was the market-discipline approach, along with other existing

protections, insufficient?

Under a market-discipline approach, the regulator’s job is to ensure that the

private sector exercises the type of diligence that enables markets to work efficiently.94

For this failure, this article examines the following hypotheses:

Hypothesis: For reasons already discussed in this article, certain foundations of a market-discipline approach have rotted.

Regulators implement a market-discipline approach by ensuring that market

participants have access to adequate information about risks and by arranging incentives

so that those who influence an institution’s behavior will suffer if that behavior generates

losses.95 In the recent financial crisis, however, disclosure inadequately conveyed

94 Cf. Ben S. Bernanke, Chairman, Board of Governors, U.S. Federal Reserve System, Remarks at the Federal Reserve Bank of Atlanta’s 2006 Financial Markets Conference, Sea Island, Georgia (May 16, 2006) (transcript available at http://www.federalreserve.gov/Boarddocs/speeches/2006/200605162/default.htm), at 6 (observing that, to the extent hedge funds are regulated solely through market discipline, government’s “primary task is to guard against a return of the weak market discipline that left major market participants overly vulnerable to market shocks”). 95 See id. Cf. Ben S. Bernanke, Chairman, Board of Governors, U.S. Federal Reserve System, Remarks at the New York University Law School, New York, New York (Apr. 11, 2007) (transcript available at


http://www.federalreserve.gov/Boarddocs/speeches/2006/200605162/default.htm

26

information about the risks for various reasons,96 including possibly that certain of the

structured finance transactions were too complex to be adequately disclosed.97

Furthermore, the incentives of managers did not appear to be fully aligned with those of

their institutions, so managers would not necessarily suffer—and, more importantly, they

would not expect to suffer—if their behavior generated losses to their institutions.98

Additionally, in the context of systemic risk, there are fundamental misalignments

between institutional and financial market interests.99 Market discipline alone is therefore

an insufficient approach.

Hypothesis: A market-discipline approach failed for other reasons.

One such reason might be the simple human greed of market participants.100 Until

recently, it appeared that a market-discipline approach worked well for the banking and

securities-brokerage industries, which in large part have been subject to this regulatory

http://www.federalreserve.gov/boardDocs/speeches/2007/20070411/default.htm) (observing that “[r]eceivership rules that make clear that investors will take losses when a bank becomes insolvent should increase the perceived risk of loss and thus also increase market discipline” and that, in “the United States, the banking authorities have ensured that, in virtually all cases, shareholders bear losses when a bank fails”). 96 See generally Part III.A, supra. 97 See supra notes 37-39 and accompanying text. 98 See supra notes 42-47 and accompanying text (observing potential agency-cost conflicts between investment bankers who structured, sold, or invested in securities and the institutions for which they worked). 99 See supra notes 68-70 and accompanying text (arguing that structured finance may have dispersed subprime mortgage risk so widely that there was no clear incentive for any given investor to monitor it). See also infra note 103 and accompanying text (observing that from the standpoint of systemic risk, a market-discipline approach is inherently suspect because no firm has sufficient incentive to limit its risk taking in order to reduce the danger of systemic contagion for other firms). 100 Cf. Roberta Romano, A Thumbnail Sketch of Derivative Securities and Their Regulation, 55 MD. L. REV. 1, 79 (1998) (discussing greed as a central factor that, in the hedge-fund context, transforms a successful hedging or moderately risky investment strategy into one of high-risk speculation). Bernanke suggests, however, a possible alternative psychological explanation, at least in the case of the failure of market-discipline in the case of LTCM’s investors: that those “[i]nvestors, perhaps awed by the reputations of LTCM’s principals, did not ask sufficiently tough questions about the risks


http://www.federalreserve.gov/boardDocs/speeches/2007/20070411/default.htm

27

approach.101 Did something change to increase the potential for greed? Query also

whether “greed” is the appropriate term, or whether this hypothesis turns on the desire by

market participants to increase risk-adjusted performance, which itself may be motivated

by (among other things) greed.

Another such reason is that, absent prescriptive rules, market discipline is

undermined by the availability heuristic102 as well as the almost endemic shortage of

funding for regulatory monitoring.

Hypothesis: At least regarding systemic risk, market discipline is inherently suspect because no firm has sufficient incentive to limit its risk taking in order to reduce the danger of systemic contagion for other firms.

Recall that the externalities of systemic failure include social costs that can extend

far beyond market participants, resulting in a type of tragedy of the commons.103 Thus, a

firm that exercises market discipline by reducing its leverage will marginally reduce the

overall potential for systemic risk; but if other firms do not also reduce their leverage, the

first firm will likely lose net asset value relative to the other firms.104

Summary. The discussion above shows that a market-discipline approach must be

supplemented and that market discipline is particularly suspect as a protection against

systemic risk.

E. Why did the rating agencies fail to anticipate the downgrades?

that were being taken to generate the high returns.” Bernanke, supra note 94, at 1. Compare the “over-reliance” hypothesis, supra note 25 and accompanying text. 101 See, e.g., Albert J. Boro, Jr., Comment, Banking Disclosure Regimes for Regulating Speculative Behavior, 74 CAL. L. REV. 431, 471 (1986); Helen A. Garten, Banking on the Market: Relying on Depositors to Control Bank Risks, 4 YALE J. ON REG. 129, 129–30, 129 n.1 (1986). 102 See supra notes 34-35 and accompanying text. 103 See supra notes 49-51 and accompanying text. 104 E-mail from Bookstaber, supra note 23.


28

This failure is particularly problematic to the extent of investor over-reliance on

rating-agency ratings.105 For this failure, this article examines the following hypotheses:

Hypothesis: Rating agencies failed because of conflicts of interest in the way they are paid.

Rating agencies are customarily paid by the issuer of securities, but investors rely

heavily on their ratings.106 This is technically a conflict, but it is not usually a material

conflict. Ratings, for example, are made independently of the fee received.107

Furthermore, the reputational cost of a bad rating usually far exceeds the income received

by giving the rating.108

In the subprime crisis, though, the conflict would have been more material than

normal because ratings were given to innumerable issuances of CDO and ABS CDO

securities, each issuance (and rating) earning a separate fee. Assuming arguendo this

created a material conflict, there is no easy solution. The question of who pays for a

rating is difficult. Historically, rating agencies made their money by selling subscriptions,

but that may not generate sufficient revenue to allow rating agencies to hire the top-flight

analysts needed to rate complex deals.109 And even if there were an easy way to get

investors to pay for ratings, that might create the opposite incentive: to err on the side of

low ratings in order to increase the rate of return to investors—thereby increasing the cost

of credit to companies.110

105 See supra notes 27-29 and accompanying text. 106 Steven L. Schwarcz, Private Ordering of Public Markets: The Rating Agency Paradox, 2002 U. ILLINOIS L. REV. 1, 3, 15. 107 Id. at 16. 108 Id. at 14. 109 Cf. id. at 16 n. 94. 110 To the extent ratings affect not only new investors but also existing investors, this analysis is complicated by the inherent conflict between those two sets of investors. Cf. Steven L. Schwarcz, Temporal Perspectives: Resolving the Conflict Between Current and Future Investors, 89 MINN. L. REV. 1044 (2005).


29

Hypothesis: Rating agencies failed to foresee that the depth of the fall of the housing market could, and indeed did, exceed their worst-case modeled scenarios?

This hypothesis provides an obvious explanation, but it begs the question of

whether the rating agency models were reasonable, at least when viewed ex ante. That

question is, effectively, identical to the earlier question of whether the failure by investors

to envision the actual worst-case scenario may have reflected, to some extent, a failure to

take a sufficiently long view of risk.111 The earlier analysis proposed two possible

answers: that the failure simply reflected a failed judgment call, made ex ante, of what

the worst-case could be like; and that the failure also may have reflected behavioral bias

caused by the availability heuristic.

It is unlikely that the failure of rating-agency models reflected behavioral bias,

since these models are constructed by multiple trained and experienced analysts.112 To

the extent the failure reflected a failed ex ante judgment call, this type of failure may be

inevitable—even for rating agencies—because the exercise of judgment involves an

inherent risk of error.113 The hope is that rating agencies, through their institutional

memory, will learn from experience and exercise better judgment in the future.

It also is possible that the rating-agency models may have failed because of fraud

in the borrower-income data.114 To this extent, rating agencies may be stymied because

they have little alternative in most cases but to accept as true the data they receive.115

Hypothesis: Rating agencies failed to fully appreciate the correlation in subprime mortgage loans when analyzing CDOs, especially ABS CDOs.

111 See supra notes 21-24 and accompanying text. 112 [cite] But cf. Gerry McNamara & Paul Vaaler, “A Management Research Perspective on How and Why Credit Assessors ‘Get it Wrong’ When Judging Borrowers” (undated draft, on file with author, suggesting that rating-agency models may have failed in part because of systematic biases resulting from behavioral factors). 113 [cite] 114 See supra note 54. 115 Schwarcz, supra note 106, at 6 (observing that rating agencies do not, and cannot pragmatically, rate for fraud).


30

Early CDOs and ABS CDOs were highly diversified.116 Later CDOs and ABS

CDOs were still diversified but were more susceptible to a finance-based link in which

prices of the underlying assets start to move in lockstep as investors hedge their exposure

to those assets.117 Furthermore, even though later ABS CDOs had significant

diversification in the ABS and MBS securities included therein, there was an underlying

correlation in the subprime mortgage loans backing the different MBS securities. Rating

agencies, however, continued to use historical cash-flow models which did not reflect the

price convergence or correlation of loans.118 [Test this hypothesis.119]

Summary. Rating agencies obviously failed to anticipate the worst-case scenario

represented by the subprime meltdown. Although this failure might have resulted in part

from conflicts of interest in the way rating agencies are paid, that is unlikely since

payment is independent of the rating and the reputational cost of issuing bad ratings

usually far exceeds the payment received. In any event, there is no easy solution to the

dilemma of how rating agencies can be paid without creating conflicts with either issuers

or investors.

A more likely explanation for the failure is that ratings are judgment calls by

human beings, and mistakes inevitably will be made.120 One might argue that rating

agencies should be more conservative, or that government should mandate more

116 [explain and cite] 117 E-mail from Bookstaber, supra note 23 (discussing this link). 118 [cite and explain better] 119 [cite1] Another possible hypothesis is that there has been rating-agency “grade inflation.” For arguments that such grade inflation occurred, see Charles W. Calomiris, “Not (Yet) a ‘Minsky Moment,’” at 18 (Oct. 2, 2007 draft, on file with author) (arguing that “[g]rade inflation has been concentrated particularly in securitized products, where the demand is especially driven by regulated intermediaries”). However even if there grade inflation, the consequences are unclear since investors were probably not misled but simply did not care so long as the securities purchased were in fact rated investment grade. 120 Cf. Standard & Poor’s, New Actions to Strengthen Ratings (Feb. 7, 2008) (on file with author) (proposing various procedural review steps to minimize human failure in the ratings process).


31

conservative ratings, but overprotection itself has a cost. If rating agencies had used more

conservative models, requiring greater overcollateralization, those models would have

been decried as wasteful if housing prices has not collapsed.

Whatever the reasons are for the failure by rating agencies to anticipate the

downgrades, it should be noted that rating agencies may not be perfect but the idea of

rating agencies is important. Individual investors face relatively high costs to assess the

creditworthiness of complex securities. Rating agencies can make this assessment on

behalf of many individual investors, thereby achieving an economy of scale.121

IV. CONCLUSIONS

This article has suggested various insights into protecting financial markets.

Additional insight comes by recognizing that most of the causes of the discussed

anomalies and failures can be divided into three categories: (i) conflicts; (ii)

complacency; (iii) complexity.122

The first category, conflicts, is the most tractable. Once identified, conflicts can

often be managed. For example, this article has shown that the excesses of the originate-

and-distribute model can be managed by aligning the interests of mortgage lenders and

investors by requiring the former to retain a risk of loss. Some conflicts, though, may be

harder to manage in practice, such as conflicts in how rating agencies are paid.

The second category, complacency, is less tractable because solutions to

complacent behavior can require changing human nature, an obviously impossible task.

After a crisis, everyone focuses on avoiding that crisis in the future (though hopefully

121 See supra note 29. 122 I am grateful to Professor Jonathan Lipson for suggesting these categories.


32

also avoiding the all-too-human tendency to fall into the rut of fighting the “last war”123).

But bounded rationality makes the crisis fade with alacrity from perceived reality.124

The subprime mortgage crisis appears to have discredited, though, at least one

form of complacency: widespread investor obsession with securities that have no

established market and, instead, are valued by being marked-to-model.

Other forms of complacency are rational and can only be addressed through

structural changes. For example, investors will almost certainly continue to over-rely on

rating-agency ratings, so long as the cost of making independent credit investigations

remains high. If rating agencies continue to provide unreliable ratings, perhaps investors

should consider whether innovative collective-action approaches, such as collective credit

determinations, might prove more reliable.125

The third category, complexity, is least tractable.126 For example, complexity was

a central culprit responsible for the failure of disclosure in the subprime crisis, but viable

solutions appear to be second best. Even beyond disclosure, complexity is increasingly a

metaphor for the modern financial system and its potential for failure, illustrated further

by the tight coupling that causes markets to move rapidly into a crisis mode; the potential

convergence in quantitatively-constructed investment strategies; the layers inserted

between obligors on loans and other financial assets and the assets’ beneficial owners,

which make it difficult to work out underlying defaults; and the problem of adverse

selection, in which investors, uncertain which investments or counterparties are sound,

123 Systemic Risk: Examining Regulators’ Ability to Respond to Threats to the Financial System: Hearing Before the H. Committee on Financial Services, 110th Cong. (Oct. 2, 2007), available at http://www.house.gov/apps/list/hearing/financialsvcs_dem/ht1002072.shtml (statement of the author, at p. 68 of transcript). 124 Cf. supra note 36 and accompanying text (observing that investors quickly forget past finance crisis and “go for the gold”). 125 Collective approaches, though, might face potential antitrust hurdles. 126 Cf. Mandel, supra note 65, at 36 (observing that “in today’s complex and globally integrated financial markets, it’s almost impossible for regulators to plug every hole”).


33

begin to shun all investments. Solving problems of financial complexity may well be the

ultimate 21st century market goal.

These categories are broad, but they do not capture certain anomalies and failures

such as systemic risk, whose uniqueness arises from a type of tragedy of the commons.

Because the benefits of exploiting finite capital resources accrue to individual market

participants whereas the costs of exploitation are distributed among an even wider class

of persons, market participants have insufficient incentive to internalize their

externalities. Government, however, can provide solutions, such as creating a liquidity

provider of last resort to purchase securities in collapsing markets (albeit at profitable

discounts to minimize moral hazard) in order to mitigate market instability that would

lead to systemic collapse.

A final possible inquiry is to ask whether periodic financial market instabilities

are harmful or, in the long run, possibly helpful to the economy. For example, perhaps

the subprime financial crisis, or something like it, was needed to turn around the

incentive-distorting liquidity glut of the past few years?127 Financial market instabilities

are believed to be acceptable if they are “relatively limited in scope,” even if deep in their

narrow impact.128 Indeed, such instabilities “may serve as critical safety valves.”129 There

are, however, two concerns. On a distributional level, market instabilities impact people,

and in the subprime crisis many of those affected have been “low-income” individuals.130

On a more fundamental level, there is “no guarantee that the next crisis won’t spread and

turn into the Big One, which undermines the whole financial system.”131

127 Cf. Balakrishnan et al., supra note 55 (discussing the liquidity glut). 128 Michael Mandel, The Economy’s Safety Valve, BUS. WK., Oct. 22, 2007, at 34. 129 Id. at 34. 130 Mandel, supra note 128, at 36-37. That many of the affected individuals have been “low-income” individuals does not conflict with this article’s earlier observation (see text accompanying note 48, supra) that QIBs are the investors who lost the most money in the subprime crisis. Low-income individuals lost money not as investors but as foreclosed homeowners. 131 Mandel, supra note 128, at 37. Cf. Michael D. Bordo, Bruce Misrach, & Anna Schwartz, NBER Working Paper Series (No. 5371), Real Versus Pseudo-International Systemic Risk: Some Lessons From History 9 (1995) (discussing how normal market


34

expansions and contractions can turn into market crises in situations of “speculative mania”). See also Vikas Bajaj & Louise Story, Mortgage Crisis Spreads Past Subprime Loans, N.Y. TIMES, Feb. 12, 2008, at 1 (discussing the spread of the subprime crisis to other markets).


duke law school legal studies research paper …...other central banks similarly have been cutting...

Documents